The following description of the background of the invention is provided to aid in understanding the invention, but is not admitted to be, or to describe, prior art to the invention. All cited publications are incorporated by reference in their entirety.
Free phosphorus and phosphonic acids and their salts are highly charged at physiological pH and therefore frequently exhibit poor oral bioavailability, poor cell penetration and limited tissue distribution (e.g. CNS). In addition, these acids are also commonly associated with several other properties that hinder their use as drugs, including short plasma half-life due to rapid renal clearance, as well as toxicities (e.g. renal, gastrointestinal, etc.) (e.g. Antimicrob Agents Chemother 1998 May; 42(5): 1146-50). Phosphates have an additional limitation in that they are not stable in plasma as well as most tissues since they undergo rapid hydrolysis via the action of phosphatases (e.g. alkaline phosphatase, nucleotidases).
Prodrugs of phosphorus-containing compounds have been sought primarily to improve the limited oral absorption and poor cell penetration. In contrast to carboxylic acid proesters, many phosphonate and phosphate esters fail to hydrolyze in vivo, including simple alkyl esters. The most commonly used prodrug class is the acyloxyalkyl ester, which was first applied to phosphate and phosphonate compounds in 1983 by Farquhar et al., J. Pharm. Sci. 72(3):324 (1983).
Cyclic phosphonate and phosphate esters have also been described for phosphorus-containing compounds. In some cases, these compounds have been investigated as potential phosphate or phosphonate prodrugs. Hunston et al., J. Med. Chem. 27: 440-444 (1984). The numbering for these cyclic esters is shown below:

The cyclic 2′,2′-difluoro-1′,3′-propane ester is reported to be hydrolytically unstable with rapid generation of the ring-opened monoester. Starrett et al. J. Med. Chem. 37: 1857-1864 (1994).
Cyclic 3′,5′-phosphate esters of araA, araC and thioinosine have been synthesized. Meier et al., J. Med. Chem. 22: 811-815 (1979). These compounds are ring-opened through the action of phosphodiesterases which usually require one negative charge.
Cyclic 1′,3′-propanyl phosphonate and phosphate esters are reported containing a fused aryl ring, i.e. the cyclosaligenyl ester, Meier et al., Bioorg. Med. Chem. Lett. 7: 99-104 (1997). These prodrugs are reported to generate the phosphate by a “controlled, non-enzymatic mechanism[s] at physiological pH according to the designed tandem-reaction in two coupled steps”. The strategy was purportedly used to deliver d4-T monophosphate to CEM cells and CEM cells deficient in thymidine kinase infected with HIV-1 and HIV-2.
Unsubstituted cyclic 1′,3′-propanyl esters of the monophosphates of 5-fluoro-2′-deoxy-uridine (Farquhar et al., J. Med. Chem. 26: 1153 (1983)) and ara-A (Farquhar et al., J. Med. Chem. 28: 1358 (1985)) were prepared but showed no in vivo activity. In addition, cyclic 1′,3′-propanyl esters substituted with a pivaloyloxy methyloxy group at C-1′ was prepared for 5-fluoro-2′-deoxy-uridine monophosphate (5-FdUMP; (Freed et al., Biochem. Pharmac. 38: 3193 (1989); and postulated as potentially useful prodrugs by others (Biller et al., U.S. Pat. No. 5,157,027). In cells, the acyl group of these prodrugs underwent cleavage by esterases to generate an unstable hydroxyl intermediate which rapidly broke down to the free phosphate and acrolein following a β-elimination reaction as well as formaldehyde and pivalic acid.
Unsubstituted cyclic phosphoramidate esters, i.e. cyclic phosphonate and phosphate esters wherein one of the ring oxygens is replaced with an NR, are also known. For example, cyclophosphamide (CPA) is representative of a class of mustard oncolytics that utilize this prodrug moiety. Although CPA is activated primarily in the liver via a cytochrome P450-catalyzed oxidation, its biological activity is outside the liver. CPA is an effective immunosuppressive agent as well as an oncolytic agent for extrahepatic cancers because one or more of the intermediate metabolites produced following P450 activation diffuses out of the liver and into the circulation. With time the intermediate(s) enter extrahepatic tissues and are thought to undergo a β-elimination reaction to generate acrolein and the active phosphoramide mustard. Both products are reported to be cytotoxic to cells. The mustard cytotoxicity results from alkylation of DNA (or RNA). Acrolein is reported to exert its toxicity via several mechanisms, including depletion of glutathione, alkylation of DNA and proteins via a Michael reaction. In addition, acrolein produces other toxicities such as the dose-limiting bladder toxicity commonly observed with cyclophosphamide therapy. Since the toxicity of these agents often hampers their use as chemotherapy agents, numerous strategies are under investigation that are designed to enhance P450 activity in or near tumors and thereby localize the activation and antiproliferative effect of these agents to the tumor. One strategy uses retroviruses or other well known techniques for introducing genes into target tissues (e.g. Jounaidi et al., Cancer Research 58, 4391 (1998)). Other strategies include the placement of encapsulated cells engineered to produce cytochrome P450s (e.g. Lohr et al., Gene Therapy 5, 1070 (1998)) at or near the tumor.
Unsubstituted cyclic phosphoramidate esters have also been prepared as potential prodrugs of the nucleosides araA and 5-fluoro-2′-deoxyuridine (Farquhar et al., J. Med. Chem. 28, 1358 1361 (1985); J. Med. Chem. 26, 1153-1158 (1983)). The compounds were studied in a mouse model of leukemia where it was assumed that if the prodrug transformation was similar to cyclophosphamide, then these agents would be useful for treating a variety of cancers including leukemias as well as carcinomas of the colon, breast and ovary. In addition, since some of the mechanisms that account for tumor cell drug resistance entail a decrease in the enzymes used to synthesize the monophosphate, the strategy was expected to possibly be beneficial in treating drug resistant tumors. The compounds were only “marginally effective” in prolonging life span in the mouse model.
A variety of substituted 1′,3′ propanyl cyclic phosphoramidates, wherein 1′ represents the carbon alpha to the nitrogen were prepared as cyclophosphamide analogs (Zon, Progress in Med. Chem. 19, 1205 (1982)). For example, a number of 2′- and 3′-substituted proesters were prepared in order to decrease the propensity of the α,β-unsubstituted carbonyl by-product to undergo a Michael reaction. 2′-Substituents included methyl, dimethyl, bromo, trifluoromethyl, chloro, hydroxy, and methoxy whereas a variety of groups were used at the 3′-position including phenyl, methyl, trifluoromethyl, ethyl, propyl, i-propyl, and cyclohexyl. Analogs with a 3′-aryl group e.g. trans-4-phenylcyclophosphamide were “moderately effective in L1210 test system and showed no activity in vivo” G. Zu Prog. Med. Chem. 19: 205-246 (1982). A variety of 1′-substituted analogs were also prepared. In general these compounds were designed to be “pre-activated” cyclophosphamide analogs that bypass the oxidation step by already existing as a 1′-substituted analog capable of producing the final compound, e.g. hydroperoxide and 1-thioether. A series of 1′-aryl analogs were also prepared in order to enhance the oxidation potential. In contrast to the 1′-hydroperoxy analogs, the 1′-aryl compounds exhibited either no activity or very poor activity in the standard anticancer in vivo screen assay, i.e. the mouse L1210 assay. The lack of activity was postulated to arise from the steric hindrance of the phenyl and therefore limited oxidation of the prodrug. Support for this postulate was the potent activity of the acyclic phenyl keto analog which exhibited activity similar to cyclophosphamide. Luderman et al. J. Med. Chem. 29: 716 (1986).
Cyclic esters of phosphorus-containing compounds are reported in the chemical literature, however they were not tested as prodrugs in biological systems. These cyclic esters include:    [1] di and tri esters of phosphoric acids as reported in Nifantyev et al., Phosphorus, Sulfur Silicon and Related Elements, 113: 1 (1996); Wijnberg et al., EP-180276 A1;    [2] phosphorus (III) acid esters. Kryuchkov et al., Izv. Akad. Nauk SSSR, Ser. Khim. 6:1244 (1987). Some of the compounds were claimed to be useful for the asymmetric synthesis of L-Dopa precursors. Sylvain et al., DE3512781 A1;    [3] phosphoramidates. Shih et al., Bull. Inst. Chem. Acad. Sin, 41: 9 (1994); Edmundson et al., J. Chem. Res. Synop. 5: 122 (1989); and    [4] phosphonates. Neidlein et al., Heterocycles 35:1185 (1993).
Numerous phosphorus-containing compounds are known to exhibit pharmacological activity but remain far from optimal due to one or more of the above-described limitations. Some of the activities described include phosphonic acids that are useful as antihypertensives and therapy for heart failure via inhibition of NEP 24.11, phosphonic acids that are useful for treating a variety of CNS conditions (stroke, epilepsy, brain and spinal cord trauma, etc.) via binding to excitory amino acid receptors (e.g. NMDA receptor), bisphosphonic acids that are useful for treating osteoporosis, phosphonic acids that are useful as lipid lowering agents (e.g. squalene synthase inhibitors), phosphonates that are useful in treating inflammation (e.g. collagenase inhibitors), phosphonates and phosphates that are useful in treating diabetes, cancer and parasitic and viral infections.
Phosphates and phosphonates that are known to be particularly useful in glucose lowering activity and therefore are anticipated to be useful in treating diabetes are compounds that bind to the AMP site of fructose 1,6-bisphosphatase (FBPase) as described in U.S. Pat. No. 5,658,889, WO 98/39344, WO 98/39343, and WO 98/39342. Other examples of phosphorus-containing drugs include squalene synthetase inhibitor (e.g. BMS 188494).
A large class of drugs known to be active against hepatitis are generally nucleoside or nucleotide analogs that are phosphorylated inside cells to produce the biologically active triphosphate. Examples include Lamivudine (3TC) and Vidarabine (araA). In each case, the drug interferes with viral replication via the triphosphate form through either inhibition of the viral DNA polymerases or DNA chain termination. Some specificity for virus-infected cells is gained by both preferential phosphorylation of the drug by virally-encoded kinases as well as by specific inhibition of viral DNA polymerases. Nevertheless, many of the nucleoside-based drugs are associated with significant non-hepatic toxicity. For example, araA frequently produces neurological toxicity (40%) with many patients showing myalgia or a sensory neuropathy with distressing pain and abnormalities in nerve conduction and a few showing tremor, dysarthria, confusion or even coma. Lok et al., J Antimicrob. Chemotherap. 14: 93-99 (1984). In other cases, the efficacy and/or therapeutic index of nucleosides is compromised by poor phosphorylation efficiencies and therefore low levels of the biologically active triphosphate (e.g. Yamanaka et al., Antimicrob. Agents and Chemother. 43, 190 (1999)).
Phosphonic acids also show antiviral activity. In some cases the compounds are antivirals themselves (e.g. phosphonoformic acid), whereas in other cases they require phosphorylation to the disphosphate, e.g. 9-(2-phosphonylmethoxyethyl)adenine (PMEA, Adefovir). Frequently, these compounds are reported to exhibit enhanced activity due to either poor substrate activity of the corresponding nucleoside with viral kinases or because the viral nucleoside kinase which is required to convert the nucleoside to the monophosphate is down regulated viral resistance. Monophosphates and phosphonic acids, however, are difficult to deliver to virally-infected cells after oral administration due to their high charge and in the case of the monophosphate instability in plasma. In addition, these compounds often have short half-lives (e.g. PMEA, Adefovir) due in most cases to high renal clearance. In some cases, the high renal clearance can lead to nephrotoxicities or be a major limitation in diseases such as diabetes where renal function is often compromised.
Liver cancer is poorly treated with current therapies. In general, liver tumors are resistant to radiotherapy, respond poorly to chemotherapy and are characterized by a high degree of cell heterogeneity. Oncolytic nucleosides such as 5-fluoro-2′-deoxyuridine, have also shown a poor response against primary liver cancers.